# bad object 4 messages from 2009-04-22 to 2009-04-22. Participants: J. Bruce Fields, Junio C Hamano, Nicolas Pitre. Thread: https://gitlist.dev/t/19009 ## J. Bruce Fields, 2009-04-22 18:48 Subject: bad object Message-ID: <20090422184842.GC9541@fieldses.org> URL: https://gitlist.dev/e/20090422184842.GC9541%40fieldses.org ``` Just today operations started warning: error: Could not read 4e69489a0ac11a9b62a25923975bfc370a30eae5 This is Documentation/networking/baycom.txt from 518e6540831c69422faecceee8f964bd439ac9d0 (pre-2.6.13), which is easy enough to find from another repository and fix the problem. But I'm curious--how do I find out what pack file that object should be in, and what the corruption looks like? This is the third time I've found corruption in a repository on this laptop, so I'm a little suspicious of the hardware. (The second time I reported here: http://marc.info/?l=git&m=121968278319668&w=2.) --b. ``` ## Junio C Hamano, 2009-04-22 19:58 Subject: Re: bad object Message-ID: <7vbpqo31fx.fsf@gitster.siamese.dyndns.org> URL: https://gitlist.dev/e/7vbpqo31fx.fsf%40gitster.siamese.dyndns.org In-Reply-To: <20090422184842.GC9541@fieldses.org> ``` "J. Bruce Fields" writes: > Just today operations started warning: > > error: Could not read 4e69489a0ac11a9b62a25923975bfc370a30eae5 > > This is Documentation/networking/baycom.txt from > 518e6540831c69422faecceee8f964bd439ac9d0 (pre-2.6.13), which is easy > enough to find from another repository and fix the problem. > > But I'm curious--how do I find out what pack file that object should be > in, and what the corruption looks like? The former you should be able to read from either "verify-pack -v" or "show-index" output. After finding the offset for the object data in the pack, a bit of "dd seek=" would be needed to extract the data. > > This is the third time I've found corruption in a repository on this > laptop, so I'm a little suspicious of the hardware. (The second time I > reported here: http://marc.info/?l=git&m=121968278319668&w=2.) > > --b. ``` ## J. Bruce Fields, 2009-04-22 21:00 Subject: Re: bad object Message-ID: <20090422210001.GI9541@fieldses.org> URL: https://gitlist.dev/e/20090422210001.GI9541%40fieldses.org In-Reply-To: <7vbpqo31fx.fsf@gitster.siamese.dyndns.org> ``` On Wed, Apr 22, 2009 at 12:58:42PM -0700, Junio C Hamano wrote: > "J. Bruce Fields" writes: > > > Just today operations started warning: > > > > error: Could not read 4e69489a0ac11a9b62a25923975bfc370a30eae5 > > > > This is Documentation/networking/baycom.txt from > > 518e6540831c69422faecceee8f964bd439ac9d0 (pre-2.6.13), which is easy > > enough to find from another repository and fix the problem. > > > > But I'm curious--how do I find out what pack file that object should be > > in, and what the corruption looks like? > > The former you should be able to read from either "verify-pack -v" or > "show-index" output. Thanks. So: bfields@pig:~/local/linux-2.6$ git verify-pack -v .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b error: Packfile index for .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b.pack SHA1 mismatch error: packed 000000e0199ffeffffffd143feffd143feff00c0 from .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b.pack is corrupt .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b.pack: bad I don't have another copy of the pack 691dbdb94.... I guess I could run git-unpack-objects and see what it produces for the bad blob? --b. > After finding the offset for the object data in the > pack, a bit of "dd seek=" would be needed to extract the data. > > > > > This is the third time I've found corruption in a repository on this > > laptop, so I'm a little suspicious of the hardware. (The second time I > > reported here: http://marc.info/?l=git&m=121968278319668&w=2.) > > > > --b. ``` ## Nicolas Pitre, 2009-04-22 23:20 Subject: Re: bad object Message-ID: URL: https://gitlist.dev/e/alpine.LFD.2.00.0904221914190.6741%40xanadu.home In-Reply-To: <20090422210001.GI9541@fieldses.org> ``` On Wed, 22 Apr 2009, J. Bruce Fields wrote: > On Wed, Apr 22, 2009 at 12:58:42PM -0700, Junio C Hamano wrote: > > "J. Bruce Fields" writes: > > > > > Just today operations started warning: > > > > > > error: Could not read 4e69489a0ac11a9b62a25923975bfc370a30eae5 > > > > > > This is Documentation/networking/baycom.txt from > > > 518e6540831c69422faecceee8f964bd439ac9d0 (pre-2.6.13), which is easy > > > enough to find from another repository and fix the problem. > > > > > > But I'm curious--how do I find out what pack file that object should be > > > in, and what the corruption looks like? > > > > The former you should be able to read from either "verify-pack -v" or > > "show-index" output. > > Thanks. So: > > bfields@pig:~/local/linux-2.6$ git verify-pack -v .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b > error: Packfile index for .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b.pack SHA1 mismatch > error: packed 000000e0199ffeffffffd143feffd143feff00c0 from .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b.pack is corrupt > .git/objects/pack/pack-691dbdb94f881bc07a0befa45f288466bde6207b.pack: bad This is really strange and suspicious to have "000000e0199ffeffffffd143feffd143feff00c0" as a SHA1 reference. It somehow lacks quite some randomness. > I don't have another copy of the pack 691dbdb94.... I guess I could run > git-unpack-objects and see what it produces for the bad blob? It will produce nothing but some other error messages. What you can do is to manually copy the object or a pack containing that object into your repository and repack which would get rid of the error. But you might want to keep the broken pack and its index around for further analysis. Nicolas ```